Обсуждение: [HACKERS] an outdated comment for hash_seq_init.
Hello. While I looked into dynahash.c, I found that the following sentense became outdated by 5dfc198. The commit removed the only usage of hash_freeze(). > * NOTE: it is possible to use hash_seq_init/hash_seq_search without any > * worry about hash_seq_term cleanup, if the hashtable is first locked against > * further insertions by calling hash_freeze. This is used by nodeAgg.c, > * wherein it is inconvenient to track whether a scan is still open, and > * there's no possibility of further insertions after readout has begun. The attached patch removes the sentence begins with "This is used by..". regards, -- Kyotaro Horiguchi NTT Open Source Software Center diff --git a/src/backend/utils/hash/dynahash.c b/src/backend/utils/hash/dynahash.c index 12b1658..1adc5841 100644 --- a/src/backend/utils/hash/dynahash.c +++ b/src/backend/utils/hash/dynahash.c @@ -1330,9 +1330,7 @@ hash_get_num_entries(HTAB *hashp) * * NOTE: it is possible to use hash_seq_init/hash_seq_search withoutany * worry about hash_seq_term cleanup, if the hashtable is first locked against - * further insertions by calling hash_freeze. This is used by nodeAgg.c, - * wherein it is inconvenient to track whether a scan is still open, and - * there's no possibility of further insertions after readout has begun. + * further insertions by calling hash_freeze. * * NOTE: to use this with a partitioned hashtable, caller had better hold* at least shared lock on all partitions of the table throughout the scan!
On Mon, Apr 24, 2017 at 3:55 AM, Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp> wrote: > While I looked into dynahash.c, I found that the following > sentense became outdated by 5dfc198. The commit removed the only > usage of hash_freeze(). > >> * NOTE: it is possible to use hash_seq_init/hash_seq_search without any >> * worry about hash_seq_term cleanup, if the hashtable is first locked against >> * further insertions by calling hash_freeze. This is used by nodeAgg.c, >> * wherein it is inconvenient to track whether a scan is still open, and >> * there's no possibility of further insertions after readout has begun. > > The attached patch removes the sentence begins with "This is used > by..". Committed. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
At Tue, 25 Apr 2017 11:00:16 -0400, Robert Haas <robertmhaas@gmail.com> wrote in <CA+TgmobU=F1JtGZ+bL91D7sE2+cBybmJ+WMQdmTT0cyV2dvO0Q@mail.gmail.com> > On Mon, Apr 24, 2017 at 3:55 AM, Kyotaro HORIGUCHI > <horiguchi.kyotaro@lab.ntt.co.jp> wrote: > > While I looked into dynahash.c, I found that the following > > sentense became outdated by 5dfc198. The commit removed the only > > usage of hash_freeze(). > > > >> * NOTE: it is possible to use hash_seq_init/hash_seq_search without any > >> * worry about hash_seq_term cleanup, if the hashtable is first locked against > >> * further insertions by calling hash_freeze. This is used by nodeAgg.c, > >> * wherein it is inconvenient to track whether a scan is still open, and > >> * there's no possibility of further insertions after readout has begun. > > > > The attached patch removes the sentence begins with "This is used > > by..". > > Committed. Thanks. -- Kyotaro Horiguchi NTT Open Source Software Center